Viewing 15 posts - 1 through 15 (of 29 total)
Jeff Moden (1/31/2008)
February 1, 2008 at 1:21 am
I think Andy has a point there. On multiprocessor machine SQL Server might use more threads to scan the index and then merge the result sets in one data stream...
January 31, 2008 at 3:40 am
Adam,
I see your point. Maybe those 'gurus' should elevate their minds a little more. Temp tables are not proprietary per se, only the way you declare them is, and all...
October 11, 2005 at 8:43 am
I'm the one whom Gary first expressed his feelings about temp tables to, but since Gary and I had already come to an agreement (that is, I like them, he...
October 11, 2005 at 1:18 am
you're right, still it's conceptually different, e.g. table variables don't get involved in transactions.
July 26, 2005 at 6:47 am
It's not a temp table, it's a table variable
July 26, 2005 at 6:21 am
Since the whole thing seems to be revamped, here's the solution I proposed one year ago.
It's pure set-based T-SQL, no cursor or structured instructions (the If's are there only for...
July 26, 2005 at 4:10 am
Try for instance to connect to a stopped SQL Server: you'll get an ODBC driver error (which is really reporting an error from the underlying network library DBNETLIB):
July 30, 2004 at 7:23 am
And you're right too
Under some circustances, setting ansi_nulls on or off can change the execution plan (By the way, in my examples it...
July 30, 2004 at 1:02 am
Although the WHERE part of the two execution plans is similar, if we look at them in their entirety they are very different, since the originating queries are different, the...
July 29, 2004 at 8:06 am
QA uses ODBC to connect. Besides, the option 'Set ansi_nulls' should be on by installation default in its Tools/Options menu
July 29, 2004 at 5:12 am
I think that, if not otherwise specified, ansi_nulls should be considered to be ON. Although at database level it's usually OFF by default, it's set automatically ON by the connection...
July 29, 2004 at 2:16 am
It's not weird, it's... ANSI!
The ANSI standard requires that every comparison against NULL always evaluates to UNKNOWN (even when you're comparing two NULLs). This means that we have a three-valued...
July 28, 2004 at 7:38 am
Just to correct myself. In my previous post I said that none of the proposed answers seemed right, and also posted the execution plans as proof.
Now, after a good night's...
July 28, 2004 at 1:09 am
But the statements didn't specify the alias in the order by. They explicity indicated myIT.Unique_ID
July 27, 2004 at 6:49 am
Viewing 15 posts - 1 through 15 (of 29 total)